Research On Resource Allocation And Task Offloading Decision Of Vehicle Network In Mobile Edge Computing Environment

In the era of information revolution,with the influx of data and information and the increasing number of vehicles,the status of vehicles in the intelligent transportation system has become more and more important.Vehicles have gradually evolved from traditional means of transportation to smart cars with smart and interconnected computing systems.People have also put forward higher requirements for smart driving,video entertainment,and safe travel.A new type of vehicle with low latency and intensive computing needs.The application poses a great challenge to the existing Internet of Vehicles.Cloud computing,which can centrally handle a large number of computing needs,was once considered an effective solution,but it sometimes has shortcomings such as extension,high energy consumption,and large network bandwidth usage.Mobile edge computing(MEC),as a new technology,deploys MEC servers with cloud computing around base stations to provide end users with computing and communication resources to reduce the amount of computation.With the development of the Io T and chip industries,on-board units have been given more computing power,and mobile edge computing technology has now been extended to on-board networks.As a new type of efficient sharing and data dissemination technology between vehicles and between vehicles and infrastructure,edge computing in in-vehicle networks has become a new computing paradigm.In the edge computing scenario of in-vehicle networks,the high mobility of vehicles and the time-varying channel environment make vehicle users more dependent on spectrum resources.Limited spectrum resources are the key to achieving communication between vehicles and the outside world,and one of the important reasons affecting the quality of vehicle communication.one.Vehicle users will also generate a large amount of data or application requests,but their own computing power is limited,and edge servers are required to handle some tasks.Therefore,how to reasonably schedule spectrum resources in the edge computing of vehicle networks to meet the different service needs of vehicle users has become a research hotspot.In this context,this paper studies the spectrum and link resource allocation of vehicle users when there is no task offloading demand,and the decision of vehicle task offloading when there is task offloading.The main research and contributions of this paper include the following aspects:(1)The communication type and system structure of the Internet of Vehicles are analyzed,the effective link and the interference link are analyzed,and the vehicle network communication system model is constructed.(2)In order to improve the spectrum resource utilization and system throughput of the Internet of Vehicles,the focus is on the spectrum resource allocation and power control of V2V(Vehicle-to-Vehicle,V2V)links.Considering the V2 V spectrum multiplexing mode,the V2 V The link reuses the spectrum resources of the V2I(Vehicle-to-Infrastructure,V2I)link,and sets the optimization target as the V2 V link transmission rate and V2 I capacity.This problem is further formulated as a Markov Decision Process(Markov Decision Process,MDP),and the state,action and reward are designed.(3)In order to obtain the optimal spectrum resource allocation strategy,a multi-agent spectrum resource dynamic allocation scheme based on Proximal Policy Optimization(PPO)reinforcement learning algorithm is proposed.The simulation results show that the channel transmission rate and In terms of the success rate of vehicle information transmission,the algorithm in this paper has a better effect than the baseline algorithm.(4)For multi-user uninstallation,a multi-user and multi-MEC task uninstallation scenario is constructed.In order to allow users to perform optimal uninstallation according to their own conditions,a partial uninstallation strategy is proposed.(5)Considering the diversification of information transmission caused by the unpredictable lines of mobile users and edge servers,in order to reduce the delay and energy consumption and find the best offloading decision,a method based on Deep Deterministic Policy Gradient(DDPG)enhancement is proposed.Task offloading schemes for learning algorithms.The simulation results show that the proposed algorithm has better performance than the baseline algorithm in terms of effectively reducing the delay and energy consumption.